Enlarger Lens Comparison

Yesterday and today I played round trying to set up a lens so that it projected on to the front element of a microscope objective which would then send the image to the camera sensor.

I had no joy with my 10X objective so today I tried the set up with a 3 X objective. I managed to get it working with the SK 45mm lens( took a lot of adjustment to get it right)

Here the thread from another forum that I used as a model fro setting up the rig

http://www.photomacrography.net/for...asc&start=15

(Second page 3rd post)

I tried set the 45mm on the bellows to 1-1 but it was more like .8 magnification. The result After some fiddling the 45mm resolved group 6 element 3 easily and was just(!!) resolving group 4. the rig was not as accurate as I would like but I think it confirms my initial test.

PHotos:

Straight out of the camera ( but reduced for posting here)





That shot gives an idea how small an area the following 100 % crop is from. The vignetting is due to the fact the image has passed through a Microscope objective ( Not you can see shadows from the resolution lines.. my Target is glass (with chrome plating for the resolution lines) to reduce the shadows I place the target on the focusing stage wrong way round( I then flip the image in post processing)

Where did you guys get the USAF test charts?

Re Resolution target: Mine is an edmunds optics product I luckily found it on ebay The chrome on glass is not the best option but as the saying goes beggars can't be choosers. It cost just under US$100.00, new they are very expensive. I would recommend a paper target rather than the type I have. At Edmund optics you are spoiled for choice and type:
For my needs the high resolution Target is too expensive for the amount of use I would give it

http://www.edmundoptics.com/test-ta...est-targets/




In writing this post I reviewed my thinking and although I initially didn't understand you now I also wonder the same.

It is quite possible, indeed likely, my first successful play at projecting the image from the lens onto a microscope objective suffered from very poor tolerances. I was linking two tilt shift bellows together by mounting them to a 200 mm Arca rail it was quite clear that I needed a rail at least twice that length to get a much more secure mounting. Also: I could have had one or both of the two objectives slightly out of focus. I honestly think at this point in time I don't have the equipment to produce repeatable results.

I am hopeful that my macro rig is protected from external vibration.The base board is a 60Lb granite slab that sits on Sorbothane feet. I would hope it is successfully reducing vibration to the level needed for this test. I believe I can't be sure that is the case until I can set up the lens and Microscope objective securely and accurately. Ideally I need to have both bellows mounted on separate focusing rails mounted to a common base rail ( I have a very clear picture in my mind of how it could be set up). Such a set up would be much more accurate than what is available to me at the moment.

My greatest fear (in relation to this test) is that my rig is not as isolated from vibration as I assume it is.


Heres what I understand about resolution testing. As Rik on the Photomacrography forum advised me the resolution limit on my cameras sensor is around Group 6 element 4.

assumption 1: Riks comment is only relevant at 1-1 magnification Group 6 element 4 provides a frequency( Is that the right word?) of 50.8 cycles per mm.

Assumption 2: If I use a magnification of 2-1 I effectively halve the frequency to 25.4 cycles per mm. If I reduce the magnification to 1-2 then I double the frequency to 101.6 cycles per mm.


Assumption 3. Riks advice as to the sensor limit is based on the laws of physics and relates to the pixel size on the sensor..on my camera that is 4.5 µm

Are my assumptions correct?


I will look at getting the gear to ensure I can perform this test properly (repeatable) but it will take a couple of months( based on budget restrictions: I do want to buy some coins!)

Edit



That doesn't surprise me. Although the best standard Enlarger lenses from SK are "Componon-S" for macro work the HM( High modulation) and Makro Iris( enlarger lenses adapted by SK for Machine Vision) are the real performers from SK( these lenses also have the Componon S designation) they usually have muli-coatings on the lenses and are optimized for macro work but I was surprised to discover not for 1-1.

From the SK web site:
the Sk 50mm F 2.8 Makro Iris is designed for 1:20 to 1:1, optimized for -0.10
My 45 mm is designed for 1:20 to 1:1, optimized for -0.17

the 80mm and 90mm have the same optimization as the 45mm

However the site does not mention how they perform reversed... the 50mm (V38mm mount) has a front and rear mount so that it can be reversed very easily.

I tested some additional enlarger/bellows lenses, hoping to find one that performs poorly at 1x. I was trying to find one that would not outresolve the sensor.

As before, the areal images were magnified by a 10x Mitutoyo infinite-corrected objective with a Thorlabs ITL200 tube lens. Magnification was as close to 1x as I could get, but it looks like the actual magnification might have been closer to 0.98 or 0.99.

Results:

1) Schneider 80mm f/4 Componon-S -- this was the worst lens at 1x. Unlike most of the other lenses, it got better when stopped down from f/4 to f/5.6:




2) Pentax 100mm f/4 Bellows-Takumar:




3) EL-Nikkor 135mm f/5.6:




4) EL-Nikkor 105mm f/5.6 (old-style):




5) EL-Nikkor 80mm f/5.6 (old style):




6) EL-Nikkor 75mm f/4:




7) Olympus 80mm f/4 bellows lens:




8) APO Rodagon 1x 75mm f/4:




- - - - - - - -

The sensor resolves around group 6 element 4, and the only lens that did not exceed this at 1x at the center when wide-open was the Schneider Componon-S 80mm f/4. However, this lens did improve at f/5.6 so it does outresolve the sensor when stopped down.

I will continue my search for bad lenses, with some Nikon bellows lenses that are hard to attach to the Olympus bellows. The 105mm f/4 has a mediocre reputation, and the 135mm is really old (1950s, adapted rangefinder lens). I'll use the Nikon PB-4 tilt/shift bellows, which might allow testing out to 10mm off-center (this might not work?).

I have to ask for coin imaging is a high resolution target necessary? I had been looking at finding one until rik told me my cameras sensor limits. Once he told me that I was pleased I had only got the stantdard 1X resolution target. I know that the hIgh resolution target is for lenses like the PN 105mm. ( all that said I would buy one at the speed of light if it came up for sale cheaply on ebay except for the Chinese copies)

Re: Assumption 3 My camera doesn't have an anti aliasing filter.... But I am looking at getting a MFT as a second camera and that will have that filter

The modern SK lenses are designed for Machine vision cameras with very small sensors hence the 1:20 to 1:1 magnification range. For coins They should perform well an MFT camera as 1-1 magnification would only be needed for the smallest of coins ( less than 13 mm in diameter) I know there are faults in comparing sensor, but one advantage of MFT is with the lower magnification requirements an MFT camera is going to see a much deeper DOF at each aperture than an APSc FF, or medium format camera.

... as the sensor size increases the needed magnification ( so that a coin fills the sensor) increases. With increasing magnification DOF decreases. So when ensurign a coin fills the sensor area: F5.6 on a MFT will see a deeper DOF than F5.6 on a APSc camera. However The APSc camera has the potential to produce qualitatively better photos. A medium format camera can produce better photos than a Full frame. { this is not the reason I want an MFT, my reason: I want a smaller lighter camera that I can use most of my normal( non macro) lenses on}


I actually believe that for ebay selling and just sharing coin pictures online a MFT camera might well be the best bet for many coin collectors. Particular the models that have built in focus stacking

Regarding the comparison between the Printing Nikkor 105mm and the Sony FE 90mm F 2.8. Based on published lens specs the Printing Nikkor is the better of the two lenses by a very wide margin Sony in advertising some new lenses( G master line) note they are now producing lenses for 40 Plus mp sensors ... so they are now producing lenses that resolve 50 lines per mm. Isn't that less than a quarter of the PNs resolution? That said on the A7rii the Sony 90mm is a dream to use...as it was designed with that camera in mind...so there is in camera correction specific to the lens.... and it is so nice to be able to AF coins. As nice as the native lens is.. the PN will remain that most used and most important lens I have!

That number might be some sort of worst case or guaranteed value. Most of the lenses I tried resolved over 100 line pairs per mm at the center.

I was lucky to own 3 95PNs in the last few years. First one was the early model. It performed well, but inexplicably the Canon 35MP was superior in my tests. I traded that lens for my current stacking setup (other guy wanted the lens, and was good at making mechanical contraptions) but was always unhappy I got rid of it. I found an "A" version for sale at a good price, but turns out it had undescribed scratches and fungus, so I returned it, but not until I had tested it to see it was superior to the 35MP. A fairly long search later, and I found another "A" version in good shape for a good price. It just takes patience, and a bit of frustration...I saw several sell for very low prices on BIN from folks who didn't know what they had. There are a few very lucky folks out there who snapped those up quickly.

The "resolution" numbers you guys are talking about seem to be MTF10, correct? While MTF10 is indeed interesting as it says the ultimate capability of the lens, I am far more interested in MTF50, or what folks call "sharpness". In my qualitative lens tests, I look for more subtle changes in the image that affect sharpness, and when I see even a small decrease in apparent sharpness I know I am already beyond MTF50. In addition to pixel-pixel contrast I look also at overall contrast, color shifts, etc when judging lens performance. I've never thought of MTF10 as very useful in actual photography, though it is a good measure of ultimate lens performance.

I'm of course looking for a 75PN (I have never seen one for sale), as well as a 150PN, though with my setup the 150PN is really too long to be practical. Mark Goodman's tests show the 150PN to be very high resolution, higher in the center even than the 105PN, but worse in the corners (even on APS-C) at 1:1. The advantage the 150PN has (at least the older models) is the magnification optimization control, which allows sharpness to be improved over a wide mag range. Newer models don't have the control and are only optimized at 1:1.

One more point I'd like to make...I prize wide coverage more than center sharpness in a lens, but the method you guys are showing only looks at center sharpness, correct?